Avid Bioservices Inc (CDMO)

[Protector]

krug, yes and no.

Yes, because if sufficient patients where enrolled early in the clinicaltrial the # of TARGETED deaths could be reach.

But without early stop the answer is more probably NO because every one saw the hockey-stick opening of the treatment centres and the more then certain related hockey-stick curve in patient enrolment.

In hockey stick their is a sudden acceleration, possibly that happened mid 2014 for the centres opening. And if Bavituximab works as good as we expect then the bavituximab 3mg/Kg arm doesn't contribute a lot to the deaths.

If we would just take PPHM's 50% and herbst etal., 2010 approx. 10 months MOS for Docetaxel alone, then Bavi patients would have a presumed MOS of 15 months. Patients enrolled in the Bavituximab arm in say JUL 2014 are only starting to be on MOS in DEC 2015 (that is last month).

Then all depends on the following:

- when were the patients effectively enrolled (possibly the bulk will have been more in Q1/2015.

- does bavituximab make 50% improvement or will it do better (PII was 113+% improvement under Bavituximab averse dose switching conditions) and with MUCH SICKER patients (ECOG groups). Bavituximab has a latency that make its workings come to late for the very sickest patients. In the current trial this will not be the case.

- this brings us to the WINDOW of the MOS table. Just looking at the MOS value (the middle element of the MOS table which by the way is NOT an average as you wrote in your post) is only one thing. It says NOTHING on the value distribution LEFT and RIGHT. In SUNRISE this can have a VERY BIG IMPACT on the # of deaths in relation to WHEN they die (see below).


Example: Two Bavi arm MOS series

3,7,14,15,15,16,16,16,16,17,[--17--],17,17,17,17,17,17,17,18,18,19
3,3,4,4,7,8,8,10,12,15,[--17--],17,17,17,18,18,18,18,19,19,21

Both tables above have a median of 17 months. Yet the second one contributes 8 deaths in the first 10 months while the second one only 2.
So in the second case the Bavi arm contributes much more deaths to be added to the control arm one, docetaxel+placebo, where 10 is expected to be the median (again, NOT average).

So:
1) Not ALL control arm patients die after 10 months (10 is the MEDIAN)

2) The LEFT side of the Bavi median will decide how well the Bavi arm contributes to eventing.

3) Also Bavi arm patients are supposed to at least have the same median
as the SOC if Bavi would be a placebo. So the better Bavi works the more the Median value will increase BUT also the higher the values on the left. In this case the elimination of the latency handicap could do miracles.